Modi: Connected TV Targeting Needs Improvement, But Measurement Is Making Strides

SethWaltersGroupM’s 25-person advanced TV team, Modi Media, is investing three times the amount it did in connected TV this year over last, driven mostly by exploding consumer demand and growth in media availability.

To support its growing investment, Modi will use interactive video ad server Innovid to aggregate audience and measure ad views across 25 connected TV devices.

“We’ve seen massive fragmentation where marketers are challenged with measuring video across every device, app and network,” said Tal Chalozin, co-founder and CTO of Innovid. “The [technology] requirements for devices like Apple and Samsung TV are totally different than integrations to TV apps like Hulu or YouTube.”

Modi will use Innovid to quantify the incremental value to advertisers if, for instance, they want to place a buy across Fox Now, WatchESPN and Sony Crackle, since Innovid just crossed the 1,000 apps mark in terms of CTV integration.

Buyer demand has catalyzed sell-side transformation.

Broadcaster ad server FreeWheel, for instance, reported in its latest Video Monetization Report (Q4 2015) that video ad views on over-the-top devices increased 76% YoY to 22% of total video ad views versus desktop. Though desktop still maintains the largest share of overall video ad views at 40%, its growth barely registered at 0.1%.

With the pendulum swinging toward CTV making up the lion’s share of overall views, agencies like Modi, which predominantly focused on household addressable buys to date, are charting a new course. 

Seth Walters, a senior partner for interactive and connected TV at Modi Media, spoke with AdExchanger about the intricacies of smart TV investment.

AdExchanger: What do you do at Modi?

SETH WALTERS: I joined the team in June 2014, from [Mindshare] where I did a lot of advanced TV work for American Express. They had a pretty substantial branded app on MVPDs like AT&T, Verizon, DirecTV, DISH and Cablevision, and also an app on Roku, Samsung, LG, Android and Amazon Fire TV. I was seeing viewing behavior change from the brand channel on traditional MVPDs to engagement that was happening on OTT platforms where consumers were a bit more lean-forward.

We started to establish relationships with platforms direct like Roku and Xbox that were retaining video inventory on their platforms and selling it. And also the pure-play programmers who had a large enough footprint across smart TVs, gaming consoles and streaming players.

What’s most challenging?

We’ve been manually customizing campaigns for advertisers across a variety of different apps and platforms to aggregate scale. There’s still a ton of fragmentation, but we know there are audiences here. Our most basic need is to verify that a view in a live-stream environment on a Samsung smart TV is the same as a view in an episode of “Real Housewives” on Bravo Now on an Apple TV, which is why we’re using Innovid. It’s ensuring that when we’re looking at real-time campaign metrics, that we have an apples-to-apples comparison.

What can you do that you couldn’t before?

Seeing data in real time, we’re able to make optimizations. For instance, if we see that a campaign delivers really heavily on Roku, we’ll call up Fox and say we need to direct some of those [network app] impressions onto other platforms to inherently increase our reach. Then we can dive in deeper post-campaign to understand reach and frequency by publisher.

Do clients see CTV as an incremental value-add to their linear or addressable buys?

Some of it is focusing on test-and-learn, while some clients are trying to reach a younger audience who’s adopting over-the-top faster than the general population. They’re more likely to look at connected TV as a viable portion of their overall plan. More and more of CTV viewership is happening in the TV apps, which is a good thing.

Can you give an example?

When we’re buying from Fox or WatchESPN, we’re running in live streams of NFL games. We’re seeing impressions that run in episodes of “Empire” on Fox Now on Apple TV. And those are metrics and reporting that when we stand up to advertisers who are heavy TV spenders, are positive. It’s not web-based, short-form video. It’s long-form, broadcast-quality content that’s airing on a television screen. It’s a complement for advertisers already buying linear.

Does CTV offer the same precision in targeting as household addressable?

From where we stand today, it’s increasingly becoming more measurable and more targetable. It is on par from a measurement standpoint, meaning we are able to look at exposure from connected TV and the impact on both brand perception and purchase behavior. We’ve started to do that for some clients now, where we’re working with Experian and pixeling the media to do some post-campaign measurement to really quantify the impact.

In terms of targeting, we’re not quite at the level of addressable, where we’re able to eliminate waste to pre-define a qualified audience and eliminate waste. But given the digital backbone, we have to believe it will be solved sooner than later.

What’s holding back targeting?

The complexity of the environment with all these operating systems. Publishers themselves have to navigate that and ensure they’re ready for dynamic ad insertion. It’s not a technology issue. It’s a complexity issue. We’ll see it come from the platforms like Roku and Xbox themselves, since they have user registration information on file, both email and credit card information.

Roku’s pretty close to pulling it off, it’s just a matter of having more ad-supported publishers migrating to the Roku ad framework, and at that point they become an MVPD to us. When we’re doing our addressable campaigns looking at in-market auto intenders for [an auto brand] we can pull counts against the Roku households and, if pricing makes sense, they’ll be part of those campaigns.

The interview has been edited for clarity and length.


Popular On AdExchanger Right Now:

 

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>